We have analysed the trial results obtained in two placebo and imipramine controlled double blind studies with a new psychotropic compound. Statistical analysis as outlined in the protocol showed a rather meagre therapeutic effect of imipramine of 1.53 points difference on HAMD-17 total score in comparison to placebo. In order to increase the discriminative power of the analysis we performed a post hoc analysis selecting centres that were able to detect a difference of at least two centres points between imipramine and placebo on HAMD-17 total score at week 6 (selective centres). All other were called nonselective. The analysis revealed that there were no statistically significant differences between the two types of centres concerning patient characteristics except that the nonselective centres recruited less patients with previous good response to antidepressant treatment. There was also some difference in drop out rates in the active treatment group wich might indicate different treatment strategies in the two groups of centres that participated in this trials.